Types and Parsing

Parse Formatted Numbers

Convert number strings that contain formatting characters (commas as thousands separators) to integers. By hand, call str.replace(',', '') to strip the commas before passing to int. With pandas, chain Series.str.replace(',', '') and .astype(int) to do the same column-wide in two steps.

By hand

Loop over strings, call .replace(',', '') on each to produce a clean digit-only string, then pass to int(). The trace shows result growing with the parsed integers.

naive.py
Replay: real traced execution (multi-file project)
strings = ['1,200', '3,450', '850', '12,000', '675']
result = []
for s in strings:
    result.append(int(s.replace(',', '')))
print('RESULT:', result)

  1. strings ← ['1,200', '3,450', '850', '12,000', '675']

    1strings = ['1,200', '3,450', '850', '12,000', '675']2result = []
    values this step['1,200', '3,450', '850', '12,000', '675']strings

  2. result ← []

    1strings = ['1,200', '3,450', '850', '12,000', '675']2result = []3for s in strings:
    values this step[]result

  3. s ← '1,200'

    2result = []3for s in strings:4    result.append(int(s.replace(',', '')))
    values this step'1,200's

  4. result ← [1200]

    3for s in strings:4    result.append(int(s.replace(',', '')))5print('RESULT:', result)
    values this step[] → [1200]result

  5. s ← '3,450'

    2result = []3for s in strings:4    result.append(int(s.replace(',', '')))
    values this step'1,200' → '3,450's

  6. result ← [1200, 3450]

    3for s in strings:4    result.append(int(s.replace(',', '')))5print('RESULT:', result)
    values this step[1200] → [1200, 3450]result

  7. s ← '850'

    2result = []3for s in strings:4    result.append(int(s.replace(',', '')))
    values this step'3,450' → '850's

  8. result ← [1200, 3450, 850]

    3for s in strings:4    result.append(int(s.replace(',', '')))5print('RESULT:', result)
    values this step[1200, 3450] → [1200, 3450, 850]result

  9. s ← '12,000'

    2result = []3for s in strings:4    result.append(int(s.replace(',', '')))
    values this step'850' → '12,000's

  10. result ← [1200, 3450, 850, 12000]

    3for s in strings:4    result.append(int(s.replace(',', '')))5print('RESULT:', result)
    values this step[1200, 3450, 850] → [1200, 3450, 850, 12000]result

  11. s ← '675'

    2result = []3for s in strings:4    result.append(int(s.replace(',', '')))
    values this step'12,000' → '675's

  12. result ← [1200, 3450, 850, 12000, 675]

    3for s in strings:4    result.append(int(s.replace(',', '')))5print('RESULT:', result)
    values this step[1200, 3450, 850, 12000] → [1200, 3450, 850, 12000, 675]result

  13. for s in strings:

    2result = []3for s in strings:4    result.append(int(s.replace(',', '')))

  14. stdout ← RESULT: [1200, 3450, 850, 12000, 675]

    4    result.append(int(s.replace(',', '')))5print('RESULT:', result)
    values this stepRESULT: [1200, 3450, 850, 12000, 675]stdout

With pandas

.str.replace(',', '') strips every comma from the string column (dtype stays object). Chaining .astype(int) then converts the cleaned strings to int64. The snapshot shows the dtype change from object to int64, matching cast-numeric-strings.

library.py 
import pandas as pd
from dalib.display import set_display
set_display()

strings = ['1,200', '3,450', '850', '12,000', '675']
df = pd.DataFrame({'x': strings})
s = df['x'].str.replace(',', '').astype(int)
result = s.tolist()
print('index:', s.index.tolist())
print('dtype before:', df['x'].dtype)
print('dtype after:', s.dtype)
print('RESULT:', result)

index: [0, 1, 2, 3, 4]
dtype before: object
dtype after: int64
RESULT: [1200, 3450, 850, 12000, 675]

Implementation notes

  • .str.replace defaults to regex=False (literal replacement). Pass regex=True only when you intend a regex pattern — e.g. to strip multiple characters at once with a character class.
  • For currency strings like '$1,200.50', chain .str.replace('[$,]', '', regex=True) then .astype(float).
  • Cross-reference: cast-numeric-strings (this chapter) for converting already-clean digit strings to integers without pre-processing.